The producers of clear liquids usually regard all visible traces of solids as wastes. The methods used often involve the addition of flocculants and filter aids which contaminate the solids. The solids content tends to be low, encouraging the use of methods which remove clear liquid from a continuously fed feed suspension tank in which the solids content increases until some deleterious effect arises, necessitating the dumping of the contents of the feed suspension tank into some other device. Invariably the accumulating solids have been steadily slowing production and productivity could benefit from some device which continuously rejected concentrated solids.
In contrast thereto, the producers of finely divided solids are usually food, mining or manufacturing industries for which the solids are desired and the liquid is best recycled. Also, the solids have specifications for size and purity, often need further processing and mostly need to be obtained at high solids content as concentrates. Filter aids will, of course, contaminate the product.
A detailed recent discussion of such needs in cross flow microfiltration is given by R. Bertera, H. Steven and M. Metcalfe, The Chemical Engineer, pp. 10-14, June, 1984.
As shown in FIG. 8 of the above publication even the latest 1984 commercial Enka Membrana A.G. filter module rapidly fouled and the clarified liquid flux continued to decline when backwashed with transmembrane clarified liquid in the constant concentration cross flow (diafiltration) mode on a fine inorganic filler.
Economically, the ability to cope with strongly fouling solids without filter aids is most pressing. This fouling problem has long been recognised and the art records some attempts to substitute gas for clarified liquid during backwashing to avoid the recycle of clarified liquid to the feed suspension. Thus Japanese unexamined Patent Kokai Publication No. 53(1978)-108882 states:
"Since the filtrate is not used in the present invention for membrane reverse cleaning, the serious defect of the prior art method, that is, returning the filtrate substantially to the crude liquid is eliminated, with obvious industrial merits." PA0 "with such measure [lumen gas alone], the deposited fine particles may be removed solely by the introduction of compressed air". PA0 (i) applying the liquid suspension to the outer surface of elastic, microporous, hollow fibres within a shell or housing whereby: PA0 (ii) discharging the retained solids from the shell by applying through the fibre lumens: PA0 (i) a shell, PA0 (ii) a plurality of elastic, hollow, microporous, polymer fibres within the shell, PA0 (iii) means for supplying pressurised feed suspension to the outside of the fibres, PA0 (iv) means for withdrawing clarified liquid from the fibre lumens, PA0 (v) means for applying liquid followed by gas under pressure to the fibre lumens to effect a transmembane cleaning of the fibres, the pressure of the liquid being sufficient to stretch substantially all of the pores of the fibres and the pressure of the gas being sufficient to ensure that the gas will pass through the larger pores of the fibres to dislodge any solids retained therein and to wash the external walls of the fibres and the interior of the shell to remove all solids from the shell to an external collection point. PA0 poly(propylene), poly(4-methylpent-1-ene), co-polymers of propylene, poly(vinylidenedifluoride), poly(sulphones), poly(phenylene sulphides), poly(phenylene oxides), phenoxy resins, poly (tetrafluoroethylene) and poly(chlorotrifluoroethylene). PA0 (a) the pore size distribution removes the smallest desired particle as is required in the case but, PA0 (b) some of the pores of the fibres will pass air and all are below 10 microns in diameter, and PA0 (c) the fibre resists strongly acidic or basic cleaning solutions and is resistant to repeated heat or chemical sterilisation, and PA0 (d) the properties of the fibre allow elastic stretching of the pores when all the pores are pressure cleaned by a volume of clarified liquid at least equal to the total pore volume followed by, PA0 (e) a gaseous pore-stretching clean at a pressure sufficient to force gas through a substantial fraction of the larger pores to clean the solids off the fibre surface and the shell wall, helping to propel the solids into a means for diversion out of the system of the accumulated solids to collection ports, and PA0 (f) the elastic properties of the fibres allow a rapid recovery of pore size after stretching before return of the feed suspension so that no oversize pores pass or entrap the material being concentrated. PA0 (a) subjecting the hydrophobic fibres either prior to or during cross-flow operation to a plug flow of a wetting agent so as to reduce the surface tension of the feed to the shell below 50 (preferably 32 to 35) dynes per centimeter at which the pore will be hydrophilic and repeating the treatment at successively lengthening intervals as traces of hydrophilic feed absorb onto the hydrophobic fibres, PA0 (b) subjecting the hydrophobic fibres, either initially or after option (a) above, to sufficient feed pressure to shrink the gas bubbles retained in the pores of the fibres so as to aid their passage from the pores and to maintain sufficient solubility of the gas in the feedstock and/or the permeate.
Transmembrane gas backwashing is impossible in very finely pored filters such as reverse osmosis membranes and ultrafilters because the pressures needed to overcome surface tension are far beyond the strengths of normal hollow fibre membranes used for these purposes; wetting liquids may pass but not gases. Any gas bubbles passing through such a membrane indicate the presence of pin hole defects in the membrane. Hence this invention has no application to reverse osmosis or to true ultrafilters.
This invention is concerned with microfilters which contain larger pores than those of ultrafilters and which range from 0.01 to 10 microns. Usually, the larger of the pores are so distributed that clarified liquids are free of all visible turbidity. Turbidity involves more than particle size, obeying and arising from well known optical laws.
Early microfilters fouled quickly since they treated particles which were not suspended by Brownian motion nor diffusion but which penetrated into the pores of a similar size range as the particles in the manner of sieve blinding.
One prior art approach to solving this problem was to operate hydrophilic microfilters in a cross flow mode with clarified liquid transmembrane backwash. High cross-flow velocities required feed suspension to be to the smaller internal filtering surface of the lumen as opposed to the larger external surface of the fibre. Thus, backwash pressures had to be limited to avoid fibre crushing. The smaller filtering surface reduced output and thus this approach was frequently not a useful solution to the fouling problem.
Another prior art approach is disclosed in Japanese Patent Kokai Publication No. 53(1978)-108882 where a hollow fibre bundle in loose "candle" configuration of hydrophilic "polyvinyl alcohol (PVA)" fibres was made to writhe during long (one minute) lumenal backwashes with air. Filter "candles" of the kind described in this Japanese specification are more akin to dead-end filters than to cross-flow shell and tube filters in that they are in the form of elongated hollow pots closed at one end.
The Examples and the single claim of the above Japanese specification make it clear that the invention disclosed therein relies upon the flow of air down the lumens causing the closed ends of the fibres to oscillate or vibrate and thereby free the fine iron colloid left by the 50 parts per million feed suspension in the pressurised feed suspension pot. The Japanese specification states:
Our experience suggests that this was because the "iron colloid" was much coarser than the pores and was restricted entirely to the surface of the fibres. Other fouling materials spread over a spectrum of sizes such as crude sugar cane juice or starch waste would have been more difficult to clean.
A development of the above "candle" configuration of hydrophilic "polyvinyl alcohol (PVA) type polymer" porous hollow fibres is to be found in U.K. Patent specification No. 2,120,952. The test suspension described in this specification contained only 5 ppm of iron oxide with an average particle size of 0.6 microns and again the relationship of particle size to pore size would ensure that cleaning was not difficult to perform. The writhing of the fibre bundle was somewhat restrained by taping the fibres loosely and enclosing them in an open sleeve which avoided the tangling and fibre breakage of the earlier mentioned Japanese Kokai Publication No. 53(1978)-108882 but the gas backwash took 5 minutes.
It should be noted that the relevant prior art uses only one type of fibre namely a polyvinyl alcohol fibre which is intrinsically hydrophilic. However, had the non-wetting iron colloids been used with hydrophobic fibres such as polypropylene fibres the gas backwash would have caused pore blockage.
This gas blockage is predicted by the theory behind "bubble point" measurement. Such gas blockage is serious in practice. As far as we are aware the prior art does not disclose gas backwashing with fibres which are intrinsically hydrophobic and had it been proposed the gas blockage effect would have been encountered after gas backwash lasting from one to five minutes.
The prior art contains one other report of gas backwashing through an even less relevant "candle in the pot" configuration but using thermoset rather than thermoplastic polymers to give a chemically resistant hydrophilic filter.
This report is contained in Soviet Union Patent No. 955,986 which concerns the use of massive (250 mm long, 70 mm bore and 25 mm walls) pots made by an undisclosed process from polished hard microspheres of thermoset phenol/formaldehyde, resorcinol/formaldehyde, pyrocatechin/formaldehyde or melamine/formaldehyde. The microspheres ranged from 0.5 to 5 microns in diameter, giving pores of 0.1 to 1.6 microns. The pores were penetrated 0.3 to 0.5 mm by the fine mineral particles. While apparently adequate for the use described in the Russian specification, the apparatus is incompatible with cross-flow configuration and minimal clarified liquid backwash.
Although prior art microfilters can recover fine solids from a liquid suspension, their operation in that regard has not been commercially successful. One reason for this has been the failure to recognise that efficient and rapid removal of the fine solids from the filtering medium is complicated by the variations in the sizes of the pores, the nature and physical characteristics of both the solids and the filtering medium and the need to stretch the pores to release retained solids.
Another reason why prior art microfilters have not been commercially successful in removing fine solids is that the fibres used have had rather poor resistance to strong hydrochloric acid and sodium hydroxide which are often used for removing from the filter natural products present in feedstocks that foul the fibres of the filters.
In addition, the hollow fibres of prior art microfilters can be rapidly destroyed by hypochlorite, chlorine and hydrogen peroxide which are routinely used for sterilisation and cleaning.
Another disadvantage of the prior art is that removal of the fouling species from the filter vessel is not carried out until after the backwash cycle is completed which increases the down time of the filter.